May 2025 arXiv papers — page 97

Arsene Brice Zotsa-Ngoufack

We study the fluctuations of a stochastic epidemic model with memory of the last infections, varying infectivity, and waning immunity, as introduced in Guerin and Zotsa-Ngoufack:arXiv preprint arXiv:2505.00601. The dynamics of the epidemic model are described by a measure-valued process with respect to infection age and individual traits. The Functional Law

Frederik vom Ende, Simon Cichy

If one wants to establish optimality of a given bipartite entanglement witness, the current standard approach is to check whether it has the spanning property. Although this is not necessary for optimality, it is most often satisfied in practice, and for small enough dimensions or sufficiently structured witnesses this criterion can be checked by hand. In th

Antoine Baldacchino-Jordan, Michał Hanasz, Mateusz Ogrodnik, Dominik Wóltański

We developed a new model for the production and propagation of spectrally resolved primary and secondary Cosmic Ray (CR) nuclei elements within the framework of the Cosmic Ray Energy Spectrum (CRESP) module of the PIERNIK MHD code. We extend the algorithm to several CR nuclei and demonstrate our code's capability to model primary and secondary CR species sim

Ajay Agarwal, M. S. Santhanam

Congestion and extreme events in transportation networks are emergent phenomena with significant socio-economic implications. In this work, we study congestion and extreme event properties on real urban street (planar) networks drawn from four cities and compare it with that on a regular square grid. For dynamics, we employ three variants of random walk with

Wei Liu, Ruochen Zhou, Yiyun Deng, Yuzhen Huang

Large Reasoning Models (LRMs) have shown remarkable capabilities in solving complex problems through reinforcement learning (RL), particularly by generating long reasoning traces. However, these extended outputs often exhibit substantial redundancy, which limits the efficiency of LRMs. In this paper, we investigate RL-based approaches to promote reasoning ef

Emilio Barucci, Yuheng Lan

We deal with the optimal execution problem when the broker's goal is to reach a performance barrier avoiding a downside barrier. The performance is provided by the wealth accumulated by trading in the market, the shares detained by the broker evaluated at the market price plus a slippage cost yielding a quadratic inventory cost. Over a short horizon, this ty

Pierre-Antoine Bernard, Rafael I. Nepomechie, Gilles Parez, Eric Ragoucy

In this paper, we investigate the ground-state entanglement entropy in inhomogeneous free-boson models in one spatial dimension. We develop a powerful method to extract the leading term in the entanglement scaling, based on the analytic properties of the inhomogeneous potential. This method is applicable to a broad class of models with smooth spatial inhomog

Xin Wang, Hao Guo, Chih-Chun Chien

The Wilczek-Zee (WZ) holonomy arises in degenerate states while the Uhlmann holonomy characterizes finite-temperature topology. We investigate possible relationships between the Uhlmann phase and the scalar WZ phase, which reflects the Uhlmann and WZ holonomy respectively, in an exemplary four-level model with two doubly degenerate subspaces. Through exact s

Antonino Ficarra, Emanuele Sgroi

In this paper, we compare the index of ass-stability $\text{astab}(I)$ and the index of $\text{v}$-stability $\text{vstab}(I)$ of powers of a graded ideal $I$. We prove that $\text{astab}(I)=1\le\text{vstab}(I)$ for any graded ideal $I$ in a 2-dimensional polynomial ring, and that $\text{vstab}(I)$ can be any positive integer in this situation. Moreover, giv

David Dinucu-Jianu, Jakub Macina, Nico Daheim, Ido Hakimi

Large language models (LLMs) can transform education, but their optimization for direct question-answering often undermines effective pedagogy which requires strategically withholding answers. To mitigate this, we propose an online reinforcement learning (RL)-based alignment framework that can quickly adapt LLMs into effective tutors using simulated student-

Marino Romero, Joshua Jeishing Wen

We present a variety of new identities involving operators in the theory of wreath Macdonald polynomials. One such family of identities gives five-term relations, analogous to the one given by Garsia and Mellit for the modified Macdonald polynomials. As a consequence, we generate tableau formulas for wreath Macdonald Pieri coefficients, which give an incredi

Markus L. Schmid

For a terminal alphabet $\Sigma$ and an attribute alphabet $\Gamma$, a $(\Sigma, \Gamma)$-extractor is a function that maps every string over $\Sigma$ to a table with a column per attribute and with sets of positions of $w$ as cell entries. This rather general information extraction framework extends the well-known document spanner framework, which has inten

Mary Luz Rodiño Montoya, Natalia A. Viana Bedoya, Carlos Henao

Given an evolution algebra associated to a connected finite graph $\Gamma$, we exhibit a free action of the group of symmetries of $\Gamma$ on the set of automorphisms of the algebra. This allows us to explicitly describe this set and we prove that a sufficient condition for it to be finite is that every automorphism is induced by a graph symmetry. Consequen

Jonathan L. Langentepe-Kong, Manasi Pranav, Safa Shoaee, Dieter Neher

The power conversion efficiency of organic solar cells has recently improved beyond 20%. The active layers of these devices comprise of at least two organic semiconductors, forming a type II heterojunction. Hereby, the device performance is determined by the kinetic interplay of various species, including localized excitons, charge transfer states as well as

Patrick Cheridito, Jean-Loup Dupret, Donatien Hainaut

In this paper, we introduce a model-based deep-learning approach to solve finite-horizon continuous-time stochastic control problems with jumps. We iteratively train two neural networks: one to represent the optimal policy and the other to approximate the value function. Leveraging a continuous-time version of the dynamic programming principle, we derive two

Ryan Bignell, Gert Aarts, Chris Allton, M. Naeem Anwar

The FASTSUM collaboration has a long-standing project examining hadronic properties using anisotropic lattice QCD. We determine the spectral properties of bottomonia at finite temperature using lattice NRQCD and describe how our newer simulations improve our control over systematic errors. Motivated by these efforts, the temperature dependence of charm hadro

Anna B. M. Souza, Clebson Cruz, Marcelo A. Moret

Quantum Machine Learning is a new computational tool that combines the quantum properties from quantum computing with the pattern recognition from machine learning. In this paper, we apply the Variational Quantum Classifier algorithm to the problem of pulsar classification of candidates from the High Time Resolution Universe 2 dataset. We use Qiskit Machine

Aniket Basak, Rajeet Ghosh, Rohit Sarma Sarkar, Chandan Goswami

We propose a Ternary Fully Device-Independent Quantum Key Distribution (TFDIQKD) protocol based on the two-party Impossible Colouring pseudo-telepathy game, utilizing maximally entangled qutrit states to enable secure key generation between distant parties. The protocol harnesses Bell inequality violations that arise from contextuality in the Kochen-Specker

Joanna Ko

Riehl and Verity have established that for a quasi-category $A$ that admits limits, and a homotopy coherent monad on $A$ which does not preserve limits, the Eilenberg-Moore object still admits limits; this can be interpreted as a completeness result involving lax morphisms. We generalise their result to different models for $(\infty, 1)$-categories, with an

Yasuyuki Kusuda

This study examines the dynamics of customer behavior with trial and return options in omnichannel retailing, where retailers face challenges in integrating physical and online stores. Recently, major retailers have begun offering customers the option of trying eligible items for a set period and returning unwanted products free of charge. However, existing

Xinyi Lu, Aditya Mahesh, Zejia Shen, Mitchell Dudley

This project examines the prospect of using AI-generated feedback as suggestions to expedite and enhance human instructors' feedback provision. In particular, we focus on understanding the teaching assistants' perspectives on the quality of AI-generated feedback and how they may or may not utilize AI feedback in their own workflows. We situate our work in a

Sébastien Piérard, Adrien Deliège, Anaïs Halin, Marc Van Droogenbroeck

Frequently, multiple entities (methods, algorithms, procedures, solutions, etc.) can be developed for a common task and applied across various domains that differ in the distribution of scenarios encountered. For example, in computer vision, the input data provided to image analysis methods depend on the type of sensor used, its location, and the scene conte

Yury Belousov, Brian Pulfer, Vitaliy Kinakh, Slava Voloshynovskiy

While foundation models demonstrate impressive performance across various tasks, they remain vulnerable to adversarial inputs. Current research explores various approaches to enhance model robustness, with Diffusion Denoised Smoothing emerging as a particularly promising technique. This method employs a pretrained diffusion model to preprocess inputs before

Patrice Salzenstein, Blandine Guichardaz, Aya Maroua Bessou, Ekaterina Pavlyuchenko

Smart contact lenses are at the forefront of integrating microelectronics, biomedical engineering, and optics into wearable technologies. This work addresses a key obstacle in their development: achieving autonomous power without compromising safety or miniaturization. We examine energy harvesting strategies using intrinsic ocular sources-particularly tear s

Niccolo Avogaro, Thomas Frick, Yagmur G. Cinar, Daniel Caraballo

Large-scale pretrained vision backbones have transformed computer vision by providing powerful feature extractors that enable various downstream tasks, including training-free approaches like visual prompting for semantic segmentation. Despite their success in generic scenarios, these models often fall short when applied to specialized technical domains wher

David C. Nak, Jasper Riebesehl, Andreas Hemmerich

Narrow-linewidth, compact and robust laser systems are of high interest, especially with prospects for portable quantum metrological applications. We realized an external cavity DFB-laser incorporating an intra-cavity EOM in a fully fiber-coupled manner. Additional active feedback on the EOM, by referencing a high finesse cavity, enabled us to reach low kHz-

Nils Bosbach, Rebecca Pelke, Niko Zurstraßen, Jan Henrik Weinstock

In today's technology-driven world, early-stage software development and testing are crucial. Virtual Platforms (VPs) have become indispensable tools for this purpose as they serve as a platform to execute and debug the unmodified target software at an early design stage. With the increasing complexity of software, especially in areas like Artificial Intelli

Carlos Stein Brito, Daniel McNamee

Deploying learned control policies in real-world environments poses a fundamental challenge. When system dynamics change unexpectedly, performance degrades until models are retrained on new data. We introduce Reflexive World Models (RWM), a dual control framework that uses world model predictions as implicit reference trajectories for rapid adaptation. Our m

Zhiyu Xu, Qingliang Chen

Since the beginning of this century, the significant advancements in artificial intelligence and neural networks have offered the potential to bring new transformations to short-term earthquake prediction research. However, currently, there is no widely used benchmark for this task. To address this, we have built a new benchmark (EPBench), which is, to our k

Yunjia Xi, Jianghao Lin, Menghui Zhu, Yongzhao Xiao

Retrieval-Augmented Generation (RAG) enhances large language models (LLMs) by grounding responses with retrieved information. As an emerging paradigm, Agentic RAG further enhances this process by introducing autonomous LLM agents into the information seeking process. However, existing benchmarks fall short in evaluating such systems, as they are confined to

Timm Spork, Christel Baier, Joost-Pieter Katoen, Sascha Klüppelholz

We introduce $(\varepsilon, \delta)$-bisimulation, a novel type of approximate probabilistic bisimulation for continuous-time Markov chains. In contrast to related notions, $(\varepsilon, \delta)$-bisimulation allows the use of different tolerances for the transition probabilities ($\varepsilon$, additive) and total exit rates ($\delta$, multiplicative) of s

Giovanni Bellettini, Shokhrukh Yu. Kholmatov

We prove a $C^{1,1}$-regularity of minimizers of the functional $$ \int_I \sqrt{1+|Du|^2} + \int_I |u-g|ds,\quad u\in BV(I), $$ provided $I\subset\mathbb{R}$ is a bounded open interval and $\|g\|_\infty$ is sufficiently small, thus partially establishing a De Giorgi conjecture in dimension one and codimension one. We also extend our result to a suitable anis

Haocheng Ju, Bin Dong

Mathematical Information Retrieval (MIR) is the task of retrieving information from mathematical documents and plays a key role in various applications, including theorem search in mathematical libraries, answer retrieval on math forums, and premise selection in automated theorem proving. However, a unified benchmark for evaluating these diverse retrieval ta

Yongjun Chen, Liping Zhang

This paper investigates the eigenvalue computation problem of the dual quaternion Hermitian matrix closely related to multi-agent group control. Recently, power method was proposed by Cui and Qi in Journal of Scientific Computing, 100 (2024) to solve such problem. Recognizing that the convergence rate of power method is slow due to its dependence on the eige

Ankita Pal, Pampa Paul

Let $G = SO_0(2,m),$ the connected component of the Lie group $SO(2,m);\ K = SO(2) \times SO(m),$ a maximal compact subgroup of $G;$ and $\theta$ be the associated Cartan involution of $G.$ Let $X = G/K,\ \frak{g}_0$ be the Lie algebra of $G$ and $\frak{g} = \frak{g}_0^\mathbb{C}.$ In this article, we have considered the special cycles associated with all po

LHCb collaboration, R. Aaij, A. S. W. Abdelmotteleb, C. Abellan Beteta

The first dedicated $Z$-boson mass measurement at the LHC with $Z \to \mu^+\mu^-$ decays is reported. The dataset uses proton-proton collisions at a centre-of-mass energy of $13$ TeV, recorded in 2016 by the LHCb experiment, and corresponds to an integrated luminosity of $1.7$ fb$^{-1}$. A template fit to the $\mu^+\mu^-$ mass distribution yields the followi

Hua Li, Shijie Lian, Zhiyuan Li, Runmin Cong

With recent breakthroughs in large-scale modeling, the Segment Anything Model (SAM) has demonstrated significant potential in a variety of visual applications. However, due to the lack of underwater domain expertise, SAM and its variants face performance limitations in end-to-end underwater instance segmentation tasks, while their higher computational requir

J. van Roestel, K. Burdge, I. Caiazzo, T. Kupfer

In a systematic search for eclipsing white dwarfs using Zwicky transient facility (ZTF) data, we found seven eclipsing double white dwarfs with orbital periods ranging from 45 minutes to 3 hours. We collected high-speed light curves, archival multi-wavelength data, and optical spectra for all systems and determined the binary parameters for each of them. We

Rui Wang, Renyu Zhu, Minmin Lin, Runze Wu

Confidence estimation is crucial for reflecting the reliability of large language models (LLMs), particularly in the widely used closed-source models. Utilizing data augmentation for confidence estimation is viable, but discussions focus on specific augmentation techniques, limiting its potential. We study the impact of different data augmentation methods on

Subrata Biswas, Mohammad Nur Hossain Khan, Bashima Islam

Multimodal question answering (QA) often requires identifying which video, audio, or sensor tokens are relevant to the question. Yet modality disagreements are common: off-camera speech, background noise, or motion outside the field of view often mislead fusion models that weight all streams equally. We present RAVEN, a unified QA architecture whose core is

Ziming Liu

A conjecture proposed by Gaetz and Gao asserts that the Cayley graph of any Coxeter group satisfies the strong hull property. In this paper, we prove this conjecture for all affine irreducible Coxeter groups of rank 3. Our approach exploits the geometry of Coxeter complexes to reduce the analysis of convex hulls to finitely many manageable configurations.

Hossein Zakerinia, Jonathan Scott, Christoph H. Lampert

Personalized federated learning has emerged as a popular approach to training on devices holding statistically heterogeneous data, known as clients. However, most existing approaches require a client to have labeled data for training or finetuning in order to obtain their own personalized model. In this paper we address this by proposing FLowDUP, a novel met

Charles Bertucci, Jean-Michel Lasry, Pierre Louis Lions

Considering a simple economy, we derive a new Hamilton-Jacobi equation which is satisfied by the value of a ''bubble'' asset. We then show, by providing a rigorous mathematical analysis of this equation, that a unique non-zero stable solution exists under certain assumptions. The economic interpretation of this result is that, if the bubble asset can produce

Johannes Branahl

Quantum Field Theory (QFT), the foundational framework of particle physics, has long existed in a state of tension between empirical success and mathematical rigor. Conventional QFT (CQFT), which underpins the Standard Model, offers unparalleled predictive accuracy but relies on inconsistent and ad hoc methods. In contrast, axiomatic QFT (AxQFT) aspires to a

Xin Huang, Ruibin Li, Tong Jia, Wei Zheng

Vision-Language Models (VLMs) are essential for multimodal tasks, especially compositional reasoning (CR) tasks, which require distinguishing fine-grained semantic differences between visual and textual embeddings. However, existing methods primarily fine-tune the model by generating text-based hard negative samples, neglecting the importance of image-based

Katsunori Fujie

Finite free convolutions, $\boxplus_d$ and $\boxtimes_d$, are binary operations on polynomials of degree $d$ that are central to finite free probability, a developing field at the intersection of free probability and the geometry of polynomials. Motivated by established regularities in free probability, this paper investigates analogous regularities for fini

Yvon Bossut

In this work we study some examples of groups definable and type-definable in NSOP1 theories. We exhibit some behaviors of these groups that differ from the ones of simple groups. We take interest in the notions of generics and stabilizers, and define the Kim-stabilizer. We apply the notion of Kim-stabilizer and the stabilizer from Hrushovsky to the context

Wangyang Ying, Haoyue Bai, Nanxu Gong, Xinyuan Wang

The data-to-equation (Data2Eqn) task aims to discover interpretable mathematical equations that map observed values to labels, offering physical insights and broad applicability across academic and industrial domains. Genetic programming and traditional deep learning-based approaches suffer from search inefficiency and poor generalization on small task-speci

Changyuan Zhao, Ruichen Zhang, Jiacheng Wang, Dusit Niyato

This paper introduces a two-stage generative AI (GenAI) framework tailored for temporal spectrum cartography in low-altitude economy networks (LAENets). LAENets, characterized by diverse aerial devices such as UAVs, rely heavily on wireless communication technologies while facing challenges, including spectrum congestion and dynamic environmental interferenc

Mendrit Latifi, Kimet Jusufi

We investigate the quantum dynamics of a charged scalar field in the near-horizon region of a near-extremal charged BTZ black hole. A controlled expansion of the Einstein-Maxwell equations reveals an emergent warped AdS$_2 \times S^1$ throat geometry threaded by a constant electric field--an ideal setting for studying Schwinger pair production, Hawking radia

Marko Tuononen, Heikki Penttinen, Duy Vu, Dani Korpi

Neural network-based radio receivers are expected to play a key role in future wireless systems, making reliable Out-Of-Distribution (OOD) detection essential. We propose a post-hoc, layerwise OOD framework based on channelwise feature aggregation that avoids classwise statistics--critical for multi-label soft-bit outputs with astronomically many classes. Re

Rinat Kashaev, Vladimir Mangazeev

We show that the exterior algebra of a vector space $V$ of dimension greater than one admits a one-parameter family of braided Hopf algebra structures, arising from its identification with a Nichols algebra. We explicitly compute the structure constants with respect to a natural set-theoretic basis. A one-parameter family of diagonal automorphisms exists, wh

Matthias Grundmann, Hannes Hartenstein

Payment channel networks are an approach to improve the scalability of blockchain-based cryptocurrencies. The Lightning Network is a payment channel network built for Bitcoin that is already used in practice. Because the Lightning Network is used for transfer of financial value, its security in the presence of adversarial participants should be verified. The

Matheus H. Martins Costa, Flavio S. Nogueira, Jeroen van den Brink

By formulating the renormalization group as a quantum channel acting on density matrices in Quantum Field Theories (QFTs), we show that ground-state expectation values of observables supported on slow momentum modes can be approximated by their averages on the fixed-point theories to which the QFTs flow. This is done by studying the fidelity between ground s

Hyun Myung Jang, Kyung Yeon Jang, Song Hee Lee, Jinwoo Park

Exciton Bohr radius (a_B) and exciton binding energy (E_b) of metal halide perovskites are two prime quantities in their applications to both light-emitting diode displays and photovoltaic devices. We develop a reliable theoretical method of simultaneously finding a_B and {\epsilon}_r^c (dielectric constant) based on the net exciton energy above the bulk ban

Hossein Hassani, Soodeh Nikan, Abdallah Shami

Vision Language Models (VLMs) employed for visual question-answering (VQA) in autonomous driving often require substantial computational resources that pose a challenge for their deployment in resource-constrained vehicles. To address this challenge, we introduce TinyDrive, a lightweight yet effective VLM for multi-view VQA in driving scenarios. Our model co

Mohammad Ali, Naeemul Hassan

This research presents a novel approach to computational framing analysis, called Semantic Relations-based Unsupervised Framing Analysis (SUFA). SUFA leverages semantic relations and dependency parsing algorithms to identify and assess entity-centric emphasis frames in news media reports. This innovative method is derived from two studies -- qualitative and

Georg Hartl, Conrad Gstöttner, Markus Schöberl

This paper examines a broadly applicable triangular normal form for x-flat control-affine systems with two inputs. First, we show that this triangular form encompasses a wide range of established normal forms. Next, we prove that any x-flat system can be transformed into this triangular structure after a finite number of prolongations of each input. Finally,

Florin Cuconasu, Simone Filice, Guy Horowitz, Yoelle Maarek

Retrieval Augmented Generation enhances LLM accuracy by adding passages retrieved from an external corpus to the LLM prompt. This paper investigates how positional bias - the tendency of LLMs to weight information differently based on its position in the prompt - affects not only the LLM's capability to capitalize on relevant passages, but also its susceptib

Julian Oelhaf, Georg Kordowich, Changhun Kim, Paula Andrea Perez-Toro

Germany's transition to a renewable energy-based power system is reshaping grid operations, requiring advanced monitoring and control to manage decentralized generation. Machine learning (ML) has emerged as a powerful tool for power system protection, particularly for fault detection (FD) and fault line identification (FLI) in transmission grids. However, ML

Zixun Guo, Simon Dixon

Moonbeam is a transformer-based foundation model for symbolic music, pretrained on a large and diverse collection of MIDI data totaling 81.6K hours of music and 18 billion tokens. Moonbeam incorporates music-domain inductive biases by capturing both absolute and relative musical attributes through the introduction of a novel domain-knowledge-inspired tokeniz

Kaiyuan Chen, Letian Fu, David Huang, Yanxiang Zhang

Recent results suggest that very large datasets of teleoperated robot demonstrations can be used to train transformer-based models that have the potential to generalize to new scenes, robots, and tasks. However, curating, distributing, and loading large datasets of robot trajectories, which typically consist of video, textual, and numerical modalities - incl

Anna R. Flowers, Christopher T. Franck, Mickaël Binois, Chiwoo Park

Gaussian processes (GPs) furnish accurate nonlinear predictions with well-calibrated uncertainty. However, the typical GP setup has a built-in stationarity assumption, making it ill-suited for modeling data from processes with sudden changes, or "jumps" in the output variable. The "jump GP" (JGP) was developed for modeling data from such processes, combining

Ana-Maria Bucur, Marcos Zampieri, Tharindu Ranasinghe, Fabio Crestani

The increasing prevalence of mental disorders globally highlights the urgent need for effective digital screening methods that can be used in multilingual contexts. Most existing studies, however, focus on English data, overlooking critical mental health signals that may be present in non-English texts. To address this gap, we present a survey of the detecti

Xinjue Wang, Esa Ollila, Sergiy A. Vorobyov

Massive machine-type communications (mMTC) are fundamental to the Internet of Things (IoT) framework in future wireless networks, involving the connection of a vast number of devices with sporadic transmission patterns. Traditional device activity detection (AD) methods are typically developed for Gaussian noise, but their performance may deteriorate when th

Wendi Zhou, Ameer Saadat-Yazdi, Nadin Kökciyan

Critical questions are essential resources to provoke critical thinking when encountering an argumentative text. We present our system for the Critical Questions Generation (CQs-Gen) Shared Task at ArgMining 2025. Our approach leverages large language models (LLMs) with chain-of-thought prompting to generate critical questions guided by Walton's argumentatio

Angelie Kraft, Judith Simon, Sonja Schimmler

Question-answering (QA) and reading comprehension (RC) benchmarks are commonly used for assessing the capabilities of large language models (LLMs) to retrieve and reproduce knowledge. However, we demonstrate that popular QA and RC benchmarks do not cover questions about different demographics or regions in a representative way. We perform a content analysis

Yujuan Liu, Ziwen Song, Tingting Lin, Biao Tang

Laser frequency fluctuation and atomic thermal motion can lead to errors in pulse duration and detuning in cold atom interferometry, thereby reducing measurement stability and fringe contrast. To address this issue, we investigate the use of super-Gaussian pulses, which are characterized by smooth temporal profiles and centralized energy distribution, in the

Alain Deville, Yannick Deville

Users of quantum mechanics, both in physics and in the field of quantum information, are familiar with the concept of a statistical mixture as introduced by von Neumann, and with the use of a density operator in that context. A density operator may also be used in another situation, introduced by Landau, with a transient coupling between the two parts of a q

Santiago Collazo, Martín F. Mestre, Carlos R. Argüelles

Stellar streams are essential tracers of the gravitational potential of the Milky Way, with key implications to the problem of dark matter (DM) model distributions, either within or beyond phenomenological $\Lambda$CDM halos. For the first time in the literature, a DM halo model based on first physical principles such as quantum statistical mechanics and the

Adrian Arnaiz-Rodriguez, Federico Errica

After a renaissance phase in which researchers revisited the message-passing paradigm through the lens of deep learning, the graph machine learning community shifted its attention towards a deeper and practical understanding of message-passing's benefits and limitations. In this paper, we notice how the fast pace of progress around the topics of oversmoothin

Haoyue Bai, Guodong Chen, Wangyang Ying, Xinyuan Wang

Geological CO2 storage (GCS) involves injecting captured CO2 into deep subsurface formations to support climate goals. The effective management of GCS relies on adaptive injection planning to dynamically control injection rates and well pressures to balance both storage safety and efficiency. Prior literature, including numerical optimization methods and sur

Michael Sullivan, Mareike Hartmann, Alexander Koller

Although the power of LLM tool-use agents has ignited a flurry of recent research in this area, the curation of tool-use training data remains an open problem$-$especially for online RL training. Existing approaches to synthetic tool-use data generation tend to be non-interactive, and/or non-compositional. We introduce RandomWorld, a pipeline for the procedu

Luiz Maia, Gerhard Fohler

The Logical Execution Time (LET) model has deterministic properties which dramatically reduce the complexity of analyzing temporal requirements of multi-rate cause-effect chains. The configuration (length and position) of task's communication intervals directly define which task instances propagate data through the chain and affect end-to-end latencies. Sinc

Andrew Caunes, Thierry Chateau, Vincent Fremont

3D semantic segmentation plays a pivotal role in autonomous driving and road infrastructure analysis, yet state-of-the-art 3D models are prone to severe domain shift when deployed across different datasets. In this paper, we propose an Unsupervised Domain Adaptation approach where a 3D segmentation model is trained on the target dataset using pseudo-labels g

Haruki Settai, Naoya Takeishi, Takehisa Yairi

For continuous systems modeled by dynamical equations such as ODEs and SDEs, Bellman's Principle of Optimality takes the form of the Hamilton-Jacobi-Bellman (HJB) equation, which provides the theoretical target of reinforcement learning (RL). Although recent advances in RL successfully leverage this formulation, the existing methods typically assume the unde

Jiaqi Zhao, Miao Zhang, Deng Xiang, Ming Li

Existing post-training quantization methods for large language models (LLMs) offer remarkable success. However, the increasingly marginal performance gains suggest that existing quantization strategies are insufficient to support the development of more compressed models. To inspire new directions for future research, this paper introduces the concept of nul

Sergios Agapiou, Ismaël Castillo, Paul Egels

The large variety of functions encountered in nonparametric statistics, calls for methods that are flexible enough to achieve optimal or near-optimal performance over a wide variety of functional classes, such as Besov balls, as well as over a large array of loss functions. In this work, we show that a class of heavy-tailed prior distributions on basis funct

Oliver Larsson, Thijs Metsch, Cristian Klein, Erik Elmroth

Recent advancements in commodity server processors have enabled dynamic hardware-based quality-of-service (QoS) enforcement. These features have gathered increasing interest in research communities due to their versatility and wide range of applications. Thus, there exists a need to understand how scholars leverage hardware QoS enforcement in research, under

Piotr P. Goldstein

The well-known Bielinski-Khalatnikov-Lifshitz (BKL) scenario for the universe near the cosmological singularity is supplemented with a few exact results following from the BKL asymptotic of the Einstein equations: (1) The cosmological singularity is proved to be an inevitable beginning or end of the universe as described by these equations. (2) Attaining the

Qi-Qi Liang, Dong Liu, Zheng-Wen Long

The Dehnen - type dark matter density distribution model is mainly used for dwarf galaxies. In recent years, researchers have speculated that black holes may exist in this dark matter model and have given the black hole metric solutions. On this basis, this paper conducts a systematic study on the quasinormal modes of a Schwarzschild black hole in a Dehnen -

Robin Segerer

This study examines cultural value alignment in large language models (LLMs) by analyzing how Gemini, ChatGPT, and DeepSeek prioritize values from Schwartz's value framework. Using the 40-item Portrait Values Questionnaire, we assessed whether DeepSeek, trained on Chinese-language data, exhibits distinct value preferences compared to Western models. Results

Abdelghani Maddi, Ahmad Yaman Abdin, Francesco Fdp de Pretis

Scientific knowledge is a key driver of technological innovation, shaping industrial development and policy decisions worldwide. Understanding how patents incorporate scientific research is essential for assessing the role of academic discoveries in technological progress. Non-Patent References (NPRs) provide a useful indicator of this relationship by reveal

Soyoung Oh, Franklin. H. Cho, We-hyo Soe, Jisoo Yu

Among the potential spin qubit candidates, yttrium phthalocyanine double-decker (YPc$_2$) features a diamagnetic metal ion core that stabilizes the molecular structure, while its magnetic properties arise primarily from an unpaired electron (S=1/2) delocalized over the two phthalocyanine (Pc) ligands. Understanding its properties in the proximity of metal el

Wei Zeng, Chuanju Xu, Yiming Lu, Qian Wang

Spectral methods employing non-standard polynomial bases, such as M\"untz polynomials, have proven effective for accurately solving problems with solutions exhibiting low regularity, notably including sub-diffusion equations. However, due to the absence of theoretical guidance, the key parameters controlling the exponents of M\"untz polynomials are usually d

Jiayuan Wu, Zhanwang Deng, Jiang Hu, Weijie Su

We consider decentralized optimization over a compact Riemannian submanifold in a network of $n$ agents, where each agent holds a smooth, nonconvex local objective defined by its private data. The goal is to collaboratively minimize the sum of these local objective functions. In the presence of data heterogeneity across nodes, existing algorithms typically r

Jinquan Wang, Xiaojian Liao, Xuzhao Liu, Jiashun Suo

Most existing training systems focus on a single region. In contrast, we envision that cross-region training offers more flexible GPU resource allocation and yields significant potential. However, the hierarchical cluster topology and unstable networks in the cloud-edge-end (CEE) environment, a typical cross-region scenario, pose substantial challenges to bu

Michał Wichrowski, Mohsen Rezaee-Hajidehi, Jože Korelc, Martin Kronbichler

This study explores matrix-free tangent evaluations in finite-strain elasticity with the use of automatically-generated code for the quadrature-point level calculations. The code generation is done via automatic differentiation (AD) with AceGen. We compare hand-written and AD-generated codes under two computing strategies: on-the-fly evaluation and caching i

Nao Mochizuki, Yasuaki Ogawa

Bondal-Kapranov's notion of enhanced triangulated categories behaves well in the framework of localization theory, in the sense that the Verdier quotient of triangulated categories can be lifted to the Drinfeld dg quotient of pretriangulated dg categories. In this paper, we develop a parallel enhancement for Nakaoka-Palu's notion of extriangulated categories

Chang Liu

Computational Fluid Dynamics (CFD) is the main approach to analyzing flow field. However, the convergence and accuracy depend largely on mathematical models of flow, numerical methods, and time consumption. Deep learning-based analysis of flow filed provides an alternative. For the task of flow field prediction, an improved Convolutional Long Short-Term Memo

Shijin Ding, Ronghua Pan, Yi Zhu

In recent years, the global existence of classical solutions to the Cauchy problem for 2D incompressible viscous MHD equations without magnetic diffusion has been proved in \cite{Ren,TZhang}, under the assumption that initial data is close to equilibrium states with nontrivial magnetic field, and the perturbation is small in some suitable spaces, say for ins

Bowen Jiang, Chaofan Ma

Caching is crucial for system performance, but the delayed hit phenomenon, where requests queue during lengthy fetches after a cache miss, significantly degrades user-perceived latency in modern high-throughput systems. While prior works address delayed hits by estimating aggregate delay, they universally assume deterministic fetch latencies. This paper tack

Lingyu Kong, Hongzhi Zhang, Jingyuan Zhang, Jianzhao Huang

Current vision-language models (VLMs) have demonstrated remarkable capabilities across diverse video understanding applications. Designing VLMs for video inputs requires effectively modeling the temporal dimension (i.e. capturing dependencies across frames) and balancing the processing of short and long videos. Specifically, short videos demand preservation

Zane K J Hartley, Lewis A G Stuart, Andrew P French, Michael P Pound

Recent years have seen substantial improvements in the ability to generate synthetic 3D objects using AI. However, generating complex 3D objects, such as plants, remains a considerable challenge. Current generative 3D models struggle with plant generation compared to general objects, limiting their usability in plant analysis tools, which require fine detail

Mattia Sangalli, Andrea Verdini, Simone Landi, Emanuele Papini

Coronal mass ejections (CMEs) represent the most extreme solar products, showing complex and dynamic structures when detected in situ. They are often preceded by a shock and carry a magnetic cloud organised as a flux rope, surrounded and permeated by turbulent fluctuations, and whose radial size expands during propagation. We investigate the internal dynamic

Marco Menale, Giuseppe Toscani

We present a possible approach to measuring inequality in a system of coupled Fokker-Planck-type equations that describe the evolution of distribution densities for two populations interacting pairwise due to social and/or economic factors. The macroscopic dynamics of their mean values follow a Lotka-Volterra system of ordinary differential equations. Unlike

Juraj Lieskovský, Jaroslav Bušek, Tomáš Vyhlídal

We present a continuous-time equivalent to the well-known iterative linear-quadratic algorithm including an implementation of a backtracking line-search policy and a novel regularization approach based on the necessary conditions in the Riccati pass of the linear-quadratic regulator. This allows the algorithm to effectively solve trajectory optimization prob

Lang Gao, Kaiyang Wan, Wei Liu, Chenxi Wang

Bias in Large Language Models (LLMs) significantly undermines their reliability and fairness. We focus on a common form of bias: when two reference concepts in the model's concept space, such as sentiment polarities (e.g., "positive" and "negative"), are asymmetrically correlated with a third, target concept, such as a reviewing aspect, the model exhibits un

Yang-hao Chan, Jonah B. Haber, Mit H. Naik, Diana Y. Qiu

Understanding exciton-defect interactions is critical for optimizing optoelectronic and quantum information applications in many materials. However, ab initio simulations of material properties with defects are often limited to high defect density. Here, we study effects of exciton-defect interactions on optical absorption and photoluminescence spectra in mo

Guillaume E. Allemand, Matteo Giantomassi, Matthieu J. Verstraete

We study charge and heat transport from first-principles in the topological Weyl semimetal TaAs. Electron-phonon coupling matrix elements are calculated using density functional perturbation theory and used to derive the thermo-electric transport coefficients, including the electrical conductivity, Seebeck coefficient, electronic thermal conductivity and the